Figure 6.12 shows a neural circuitry model, called a Reichardt detector, which responds to directional motion in the human vision system. Neurons in the ganglion layer and LGN detect simple features in different spots in the retinal image. At higher levels, motion detection neurons exist that respond when the feature moves from one spot on the retina to another nearby spot. The motion detection neuron activates for a feature speed that depends on the difference in path lengths from its input neurons. It is also sensitive to a particular direction of motion based on the relative locations of the receptive fields of the input neurons. Due to the simplicity of the motion detector, it can be easily fooled. Figure 6.12 shows a feature moving from left to right. Suppose that a train of features moves from right to left. Based on the speed of the features and the spacing between them, the detector may inadvertently fire, causing motion to be perceived in the opposite direction. This is the basis of the wagon-wheel effect, for which a wheel with spokes or a propeller may appear to be rotating in the opposite direction, depending on the speed. The process can be further disrupted by causing eye vibrations from humming . This simulates stroboscopic conditions, which discussed in Section 6.2.2. Another point is that the motion detectors are subject to adaptation. Therefore, several illusions exist, such as the waterfall illusion  and the spiral aftereffect, in which incorrect motions are perceived due to aftereffects from sustained fixation [18,205].
Steven M LaValle 2016-12-31